CN108462663B

CN108462663B - System and method for IQ mismatch calibration and compensation

Info

Publication number: CN108462663B
Application number: CN201710916865.6A
Authority: CN
Inventors: 苟天高; 尼兰詹·拉特纳卡; 普拉纳夫·达亚尔; 根纳季·坲吉恩; 李正元
Original assignee: Samsung Electronics Co Ltd
Current assignee: Samsung Electronics Co Ltd
Priority date: 2017-02-22
Filing date: 2017-09-30
Publication date: 2022-07-15
Anticipated expiration: 2037-09-30
Also published as: US11129122B2; US20190380101A1; KR102318134B1; CN108462663A; KR20180097112A; US20180242269A1; US10412698B2; TW201832481A; US20220007318A1; TWI826357B

Abstract

A system and method for IQ mismatch calibration and compensation is disclosed. A method for providing IQ mismatch (IQMM) compensation comprising: transmitting a single frequency signal at an original frequency; determining a first response of the corrupted signal at the original frequency and a second response of the corrupted signal at the corresponding image frequency; determining an estimate of a frequency response of the compensation filter at the original frequency based on the first response and the second response; determining a snapshot of a frequency response of the compensation filter by repeating the steps of transmitting a single frequency signal, determining a first response and a second response, and determining an estimate of the frequency response of the compensation filter by sweeping the single frequency signal through a plurality of frequency steps; converting the frequency response of the compensation filter into a plurality of time domain filter taps of the compensation filter by performing a pseudo-inverse of a time-frequency transform matrix; a time delay is determined that provides the smallest LSE for the corresponding time domain filter tap.

Description

System and method for IQ mismatch calibration and compensation

This application claims benefit and priority from U.S. provisional patent application serial No. 62/461,994 filed on 2017, month 2, and 22, which is incorporated herein by reference in its entirety.

Technical Field

The present disclosure relates generally to wireless communication systems, and more particularly, to a system and method for IQ mismatch calibration and compensation.

Background

In an ideal Frequency Modulation (FM) wireless communication receiver, the analog Front End (FE) exhibits the same amplitude and phase response on the in-phase (I) and quadrature (Q) branches. However, in actual operation, mismatches and imbalances between the I-branch and the Q-branch are unavoidable due to operating conditions and imperfections caused by components of the wireless communication receiver, such as mixers, analog low-pass filters, and analog-to-digital converters (ADCs). This mismatch and imbalance introduces an image signal at the image frequency of the baseband frequency, which can interfere with the demodulation and/or modulation process of the original signal. The image signal may degrade the performance of the wireless communication receiver. To reduce IQ mismatch and imbalance, a number of IQ mismatch compensation (IQMC) techniques based on Digital Signal Processing (DSP) have been proposed.

IQ mismatch and imbalance are major causes of Radio Frequency (RF) impairment in modern direct conversion RF receivers. In a typical IQMC architecture, an adaptive filter can find filter coefficients by iteratively exploring desired properties based on the actual received signal. However, those filter coefficients obtained by an iterative process may not meet the increasing demand for high data rates in emerging wireless communication applications.

Disclosure of Invention

According to one embodiment, a method for providing IQ mismatch (IQMM) compensation comprises: transmitting a single frequency signal at an original frequency; determining a first response of the corrupted signal at the original frequency and a second response of the corrupted signal at the corresponding image frequency; determining an estimate of a frequency response of the compensation filter at the original frequency based on the first response and the second response; determining a snapshot of a frequency response of the compensation filter by repeating the steps of transmitting a single frequency signal, determining a first response and a second response, and determining an estimate of the frequency response of the compensation filter by sweeping the single frequency signal with a plurality of frequency steps having intervals corresponding to a plurality of subcarrier frequencies of the original frequency (sweep); converting the frequency response of the compensation filter into a plurality of time domain filter taps of the compensation filter by performing a pseudo-inverse of a time-frequency transform matrix; the time delay that provides the smallest time domain filter tap is determined based on a plurality of Least Squared Errors (LSEs) for the respective LSE.

According to one embodiment, a method for providing IQ mismatch (IQMM) compensation comprises: estimating filter coefficients corresponding to a plurality of filter taps of the compensation filter before reception of the normal signal based on a static calibration scheme; setting each of the plurality of filter taps using the respective estimated filter coefficients; setting an initial value of a time delay tap to zero or based on an estimated value obtained using a static calibration scheme; the filter coefficients of the time delay taps are estimated during reception of the normal signal based on an iterative scheme using an adaptive filter.

According to one embodiment, an apparatus comprises: a signal generator for generating and transmitting a single frequency signal at an original frequency; a compensator comprising a time delay and a plurality of time domain filter taps; and compensation logic to perform static calibration of the compensator. The compensation logic is configured to: determining a first response of the corrupted signal at the original frequency and a second response of the corrupted signal at the corresponding image frequency; determining an estimate of the frequency response of the compensation filter at the original frequency based on the first response and the second response; determining a snapshot of a frequency response of the compensation filter by repeating the steps of transmitting a single frequency signal, determining a first response and a second response, and determining an estimate of the frequency response of the compensation filter by sweeping the single frequency signal through a plurality of frequency steps having an interval corresponding to a plurality of subcarrier frequencies of the original frequency; converting the frequency response of the compensation filter into a plurality of time domain filter taps of the compensation filter by performing a pseudo-inverse of a time-frequency transform matrix; the time delay that produces the smallest time domain filter tap is determined based on a plurality of Least Squared Errors (LSEs) for the respective LSE.

The above and other preferred features, including various novel details of implementation and combination of events, will be more particularly described with reference to the accompanying drawings and pointed out in the claims. It will be understood that the specific systems and methods described herein are shown by way of illustration only and not as limitations. As will be understood by those skilled in the art, the principles and features described herein may be employed in various embodiments and numerous embodiments without departing from the scope of the disclosure.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate presently preferred embodiments and, together with the general description given above and the detailed description of the preferred embodiments given below, serve to explain and teach the principles described herein.

Fig. 1 illustrates an exemplary block diagram of an exemplary IQMC system in accordance with one embodiment;

FIG. 2 illustrates an exemplary diagram of an IQ mismatch model (IQMM) according to one embodiment;

FIG. 3 illustrates an exemplary plot of non-causal filter coefficients for a non-causal (non-causal) filter according to one embodiment;

FIG. 4 illustrates an exemplary plot of filter coefficients for a causal (cause) filter according to one embodiment;

FIG. 5 shows a flow diagram of a training-based calibration scheme according to one embodiment;

FIG. 6 illustrates an example real filter in accordance with one embodiment;

fig. 7 illustrates a block diagram of an example complex IQMC system, in accordance with one embodiment; and

fig. 8 shows a block diagram of an example IQMC system, according to one embodiment.

The figures are not necessarily to scale and elements of similar structure or function may be represented generally by the same reference numerals for illustrative purposes throughout the figures. The drawings are intended only to facilitate the description of the various embodiments described herein. The drawings are not intended to depict every aspect of the teachings disclosed herein nor are they intended to limit the scope of the claims.

Detailed Description

Each of the features and teachings disclosed herein may be utilized separately or in conjunction with other features or teachings to provide IQ mismatch calibration and compensation. Representative examples of the use of many of these additional features and teachings, both separately and in combination, will be described in more detail with reference to the accompanying drawings. Such detailed description is merely intended to teach a person of ordinary skill in the art additional details for practicing aspects of the present teachings and is not intended to limit the scope of the claims. Thus, combinations of features disclosed above in the detailed description may not be necessary to practice the present teachings in the broadest sense, and are instead taught merely to describe particularly representative examples of the present teachings.

In the following description, for purposes of explanation only, specific nomenclature is set forth to provide a thorough understanding of the present disclosure. However, it will be apparent to one skilled in the art that these specific details are not required in order to practice the teachings of the present disclosure.

Some portions of the detailed description herein are presented in terms of algorithms and symbolic representations of operations on data bits within a computer memory. These algorithmic descriptions and representations may be used by those skilled in the data processing arts to effectively convey the substance of their work to others skilled in the art. An algorithm is here, and generally, considered to be a self-consistent sequence of steps leading to a desired result. The steps are those requiring physical manipulations of physical quantities. Usually, though not necessarily, these quantities take the form of electrical or magnetic signals capable of being stored, transferred, combined, compared, and otherwise manipulated. It has proven convenient at times, principally for reasons of common usage, to refer to these signals as bits, values, elements, symbols, characters, terms, numbers, or the like.

It should be borne in mind, however, that all of these and similar terms are to be associated with the appropriate physical quantities and are merely convenient labels applied to these quantities. Unless specifically stated otherwise as is clear from the following description, it is understood that: throughout the specification, discussions utilizing terms such as "processing," "computing," "calculating," "determining," "displaying," or the like, refer to: the operation and processing of computer systems, or similar electronic computing devices, that manipulate and transform data represented as physical (electronic) quantities within the computer system's registers and memories into other data similarly represented as physical quantities within the computer system memories, registers or other such information storage, transmission or display devices.

The algorithms illustrated herein are not inherently related to any particular computer or other apparatus. It may prove convenient to use a variety of general purpose systems, computer servers, or personal computers with programs in accordance with the teachings herein, or to construct a more specialized apparatus to perform the required method steps. The required structure for a variety of these systems will appear from the description below. It will be appreciated that a variety of programming languages may be used to implement the teachings of the disclosure as described herein.

Furthermore, various features of the representative examples and the dependent claims may be combined in ways that are not specifically or explicitly enumerated in order to provide additional useful embodiments of the present teachings. It is also expressly noted that all value ranges or indications of groups of entities disclose every possible intermediate value or intermediate entity for the purpose of original disclosure and for the purpose of limiting claimed subject matter. It is also expressly noted that the sizes and shapes of the components shown in the figures are designed to help understand how the present teachings are implemented, and are not intended to limit the sizes and shapes shown in the examples.

The present disclosure provides an IQ mismatch compensation (IQMC) system and method that can reduce the effects of IQ mismatch and/or IQ imbalance using Digital Signal Processing (DSP) techniques. Fig. 1 illustrates an exemplary block diagram of an exemplary IQMC system in accordance with one embodiment. IQMC system 100 may be implemented as a stand-alone signal receiver or as a signal receiver integrated in a wireless communication transceiver.

In the time domain, the IQMC system 100 receives an input signal z (t) (also referred to herein as a mismatch signal or impairment signal) that includes mismatch and imbalance, and generates a signal y (t) that compensates for the input signal z (t). The input signal z (t) comprising a real part z_i(t) and imaginary part z_q(t) complex impairment signal. The IQMC system 100 includes: a filter 113, an operation block 112 to operate on the input signal z (t), a delay block 111 to be applied on the main path to the input signal z (t), and an adder block 116 to add the signals output from the delay block 111 and the filter 113.

According to one embodiment, the operation block 112 is a complex conjugate unit that obtains the complex conjugate of the input signal z (t). In another embodiment, the operation block 112 is a real unit that obtains the real part of the input signal z (t). In some embodiments, the complexity of the conjugation operation performed by the operation block 112 can be reduced by obtaining only the real part of the input signal z (t). The output of the operation block 112 is fed as an input to a filter 113.

According to one embodiment, the IQMC system 100 may comprise a complex compensator (CVC) using a complex compensation filter. For example, the input signal z (t) may be a signal comprising a real part z_i(t) and imaginary part z_q(t), and the filter 113 may be a complex compensation filter.

The output signal from the filter 113 is fed to an adder block 116 that combines the filtered signal with the delayed input signal, i.e., z (t-D), to produce a compensated signal y (t). Regardless of whether the complex conjugate of the input signal z (t) or the real part of the input signal z (t) is used, the IQMC system 100 may use CVC because the filter 113 is a complex filter.

According to one embodiment, the present IQMC system 100 may be parameterized by the number of delays D (e.g., delay 2 indicates a 2-sample delay) and the number of filter taps N in the filter 113 (N is the number of filter coefficients). In general, the optimal delay may increase as the number N of filter taps in the filter 113 increases. The present IQMC system 100 determines the filter coefficients taking into account the number N of filter taps in the filter 113 and the optimal delay of the delay block 111.

According to one embodiment, the present IQMC system 100 may provide an IQ mismatch calibration to determine the optimal delay D and the optimal filter coefficients of the filter 113. In one embodiment, the present IQMC system 100 employs a single-frequency training signal sequence having a baseband frequency to determine optimal filter coefficients for a given delay D. The baseband frequency refers to the original frequency of the received signal. The original signal can occupy a band of a certain size, for example 20 MHz. The single frequency training signal sequence can be located within a frequency range sampling discrete frequencies within the frequency range, e.g., 1MHz spacing size. The filter coefficients can be estimated by a single frequency training signal. In the time domain, the compensated signal y (t) can be represented as a function of the input signal z (t) with delay D and filter coefficients w (t):

y (t-D) + w (t) { z (t) } (equation 1)

In equation 1, a real unit is used as an example as the operation block 112, and a complex conjugate unit may be used without departing from the scope of the present disclosure.

Fig. 2 shows an exemplary diagram of an IQ mismatch model (IQMM) according to one embodiment. The iqm shown in fig. 2 is based on I/Q downconversion. The output of IQMM is fed into the IQMC system of fig. 1. IQMM 200 may demodulate and divide the received signal s (t) into an in-phase (I) signal path and a quadrature (Q) signal path. Each of the I and Q paths includes a mixer (211 and 212, respectively) and an analog filter (h)₁(t) and h₂(t)). The frequency used by the

mixers

211 and 212 is ω_LORespectively outputs a down-converted signal m in an I path and a Q path_i(t) and m_q(t) of (d). Through an analog filter h₁(t) and an analog filter h₂(t) for the down-converted signal m_i(t) and m_q(t) filtering to produce output signal z on the I and Q paths, respectively_i(t) and the output signal z_q(t) of (d). The signals on the I and Q paths may introduce mismatches and imbalances including: 1) mismatch of gain g and phase phi at mixer 211 and mixer 212, and 2) analog filter h ₁(t) and an analog filter h₁(t) overall frequency response.

The IQMM may be frequency-independent (FI) -IQMM or frequency-dependent (FD) -IQMM. The iqm shown in fig. 2 can be FI-iqm or FD-iqm. FI-IQMM may be applied to signals with non-uniform gain g ≠ 1 and non-zero phase φ ≠ 0. FD-IQMM may include an analog filter h₁(t) and an analog filter h₂(t) wherein h₁(t) is the impulse response of the analog filter along the I path, h₂(t) is the impulse response of the analog filter along the Q path.

Demodulated and divided into I and Q pathsThe received signal s (t) may comprise a desired original signal and an image signal representing a gain and/or phase imbalance introduced on the desired original signal, wherein the gain and/or phase imbalance is passed through the

mixers

211 and 212 and the analog filter h of the IQMM 200₁(t) and an analog filter h₂(t) results of signal processing. The present IQMC system attempts to minimize the influence of the image signal.

The mismatched signal z (t) of the IQMM 200 may be represented as a received signal s (t) and an image signal s^*(t) as a function of time. For example, the mismatch signal z (t) can be expressed as:

z(t)＝g₁(t)*s(t)+g₂(t)*s^*(t), (equation 2)

Wherein, the first and the second end of the pipe are connected with each other,

g₁(t) and g₂(t) denotes the signal s (t) and the image signal s for reception ^*(t) a complex scaling factor. E.g. scaling factor g₁(t) is the impulse response of the effective filter through which the original signal passes, the scaling factor g₂(t) is the impulse response of the effective filter through which the image signal introduced by the IQ mismatch passes. The mismatch signal z (t) of equation 2 can be expressed in the frequency domain using a fourier transform as an equation:

Z(f)＝G₁(f)S(f)+G₂(f)S^*(-f), (eq 4)

Wherein the content of the first and second substances,

referring to the present IQMC system shown in fig. 1, the compensated signal y (t) may be scaled by a complex scaling factor g₁(t) and g₂(t), the received signal s (t) and the image signal s^*(t) is expressed as:

from equation 6, the method can be implemented by

To obtain a signal s from which the image signal s can be completely eliminated^*(t) an optimal filter coefficient w (t). This leads to an optimal filter coefficient W in the frequency domain_OPT(f)：

In the example of iqm shown in fig. 2, the optimal filter is a non-causal filter with a large number of filter taps. The causal filter has a zero value at the negative index (index) of the filter taps. Conversely, a non-causal filter may have a non-zero value at the negative exponent of the filter taps. FIG. 3 illustrates an exemplary plot of filter coefficients for a non-causal filter according to one embodiment.

According to one embodiment, the present IQMM is based on a pole mismatch (pole mismatch) between two third order Butterworth (Butterworth) filters on the I path and the Q path. The example in fig. 3 shows filter taps of the complex optimal filter spanning from-20 to 20, where the complex optimal filter shows only the real part. The optimal filter corresponds to the IQMM model with a 40MHz third order Butterworth filter with pole mismatch [ 2%, 2%, 2% ], gain mismatch 0%, and phase mismatch of 0 degrees. The negative filter tap (negative filter tap) has a non-zero value; thus, the filter is a non-causal filter.

According to one embodimentThe optimal filter is approximated with a causal filter having a finite number of filter taps. Here, the optimal filter may refer to a causal optimal filter having a limited number of filter taps, wherein the causal optimal filter is an approximation of a non-causal optimal filter having a large number of filter taps. When the optimal filter W is approximated by a filter having a finite number of filter taps_OPTOne or more negative taps may be included. FIG. 4 illustrates an exemplary plot of filter coefficients for a causal filter according to one embodiment.

In the case of a causal filter, the optimal filter may be approximated with a Finite Impulse Response (FIR) filter by truncating the optimal filter by selecting only a limited number of filter taps N. The FIR filter has a finite number of taps. The FIR may be a non-causal filter if there is at least one negative tap. In the example shown in fig. 4, the FIR filter without delay includes filter taps 0 to 4 having high energy/amplitude, and does not include negative taps. However, it was observed that: filter tap-1 of the optimum filter has a larger amplitude than the amplitude of filter tap 4. In this case, the FIR filter may include a filter tap-1. To select filter tap-1, an additional delay may be introduced on the feed-through path (e.g., in delay block 111) such that filter taps-1 through 3 are all effectively shifted right by 1 unit to occupy filter taps 0 through 4. The unit here refers to the number of samples, and its value is the same value as that of the delay D sample introduced into the delay block 111 of fig. 1. Thus, the additional delay may improve the filtering performance of the original filter. In this case, the impairment parameters of the IQMM are assumed to be known, so that an FIR approximation of the optimal filter can be determined based on the IQMM impairment parameters. However, in practice, the impairment parameters of the IQMM may not be available a priori, and a training scheme may be used to obtain the impairment parameters.

According to one embodiment, the present IQMC system may estimate filter coefficients of a compensation filter using one or more pilot (pilot) single frequency signals. For example, the present IQMC system may use one or more pilot single frequency signals to provide a training-based IQMC calibration. The training-based IQMC calibration first transmits a single-frequency signal at a selected frequency within the frequency range of the desired signal. The present IQMC system observes and analyzes the response of the corrupted signal (i.e., the received signal without compensation) at both the original and image frequencies of the received signal. Based on the response at the original and mirror frequencies, the IQMC system may also estimate the response of the compensation filter at other frequencies. The present IQMC system can sweep a single frequency within a frequency range in a certain step size to obtain a snapshot (snapshot) of the frequency response of the compensation filter within the frequency range. The present IQMC system converts the frequency response of the compensation filter into time domain taps by performing a pseudo-inverse (pseudo-inverse) of the time-frequency transform matrix. The present IQMC system can check the frequency response of the compensation filter with the same set of single-frequency signals by applying different delays.

According to one embodiment, the IQMC system operates at a selected frequency

K continuous time single frequency signals are generated. According to one embodiment, the selected frequency

May be a multiple of subcarrier spacing. For each K e { 1.,. K }, a Discrete Time Fourier Transform (DTFT) is used, and the IQMC system uses a Discrete Time Fourier Transform (DTFT) to determine the received time-domain signal z (t) at the normalized frequency

And

lower is represented as

And

the frequency component of (a). After the received signal is processedIs denoted by f_s：

The following quantities are determined by the following equations:

for each D e {0, …, N-1}, the IQMC system determines:

by means of representation

And

this step can also be written in the form of a matrix as follows:

W_DDW' (eq 11)

The K × N discrete time fourier (DFT) matrix is denoted as F, the term of which is given by equation 12:

where K is an element { 1.,. K }, N is an element { 0.,. N-1} (equation 12)

Is the operating frequency at which the IQMC system operates.

According to one embodiment, the IQMC system pre-computes and stores the pseudo-inverse of the DFT matrix F as:

pinv(F)＝(F^HF)^-1F^H(equation 13)

The IQMC system loads the pseudo-inverse of the DFT matrix F and taps N of the optimal filterFIR approximation w_opt，N，DThe calculation is as follows:

w_opt，N，D＝pinv(F)W_Dpinv (F) DW' (equation 14)

In another embodiment, the IQMC system pre-computes pinv (f) D for each delay D and loads the product in its entirety to avoid computing the matrix product of 3 matrices pinv (f), D and W'. For each w _opt，N，DThe present IQMC system can calculate the minimum square error defined as:

LSE_D＝||W_D-Fw_opt，N，D||²(equation 15)

The present IQMC system uses the following equation to select the optimal delay D that yields the smallest Least Squared Error (LSE):

D＝arg min_D LSE_D(equation 16)

In this case, the smallest squared error between the desired frequency response and the frequency response of the designed filter is the matrix used to select the optimal delay D.

Under ideal conditions, a static calibration scheme may find all filter coefficients for a given tuple of parameters including signal bandwidth, frequency band, and frequency channel. However, in practice, the number of possible combinations of these three parameters is too large, and the number of static calibrations for each combination is greatly limited. To reduce complexity, a static calibration scheme may be performed for each possible signal bandwidth using an arbitrarily selected pair of frequency bands and frequency channels. Note that the present static calibration scheme may compensate both FI-iqm and FD-iqm for a given frequency band, channel and signal bandwidth configuration. The FI-IQMM varies with frequency band/channel, so it may not be practical to apply the present static calibration scheme for all possibilities of three parameters. A hybrid calibration scheme will be described for the following example.

According to one embodiment, the present IQMC system may perform a static calibration for each bandwidth to determine all but one filter coefficient. The one filter coefficient corresponds to the value of tap D, which is the delay value used in delay block 111 in fig. 1. In this embodiment, the present IQMC system determines the filter coefficients of the tap D, which may vary with the frequency band/channel, based on the adaptive process. The adaptive processing uses a normal received signal rather than a single frequency training signal. The adaptation process looks for the value of tap D taking into account the number of filter taps selected and the delay D. The present IQMC system may determine an improvement to the original filter coefficients obtained by the static calibration scheme based on the convergence of the remaining filter coefficients. It is known that the convergence is guaranteed for commercial wireless standards such as LTE, 3G and Wi-Fi.

FD-IQMM is not channel/band dependent because it is mainly caused by a mismatch between the two analog filters which is only bandwidth dependent and not channel or band dependent. On the other hand, the FI-IQMM is channel/band dependent because it is mainly introduced at the channel/band dependent mixer. Based on these observations, the present IQMC system separates the compensation into FI-IQMM compensation and FD-IQMM compensation.

According to one embodiment, the ratio of the analog filter

Is defined as H_d(f) Its time domain impulse response is defined as h_d(n)。W_opt(f) Time domain response w of_opt(t) can be expressed as:

for the filter tap D, it depends mainly on FI-IQMM:

for the other taps, if the filter coefficients can be obtained in the training channel by the following equation:

then the time domain response w_opt(t) may depend on both FI-IQMM and FD-IQMM. Here, the training channel pointer is for an arbitrarily selected frequency band and an arbitrarily selected channel of a particular wireless standard (e.g., LTE, 3G, and Wi-Fi). Time domain response

n ≠ D may be applied to other test channels. In this case, the actual coefficients for n ≠ D in the test channel can be given by the following equation:

thus, the error in the form of the ratio between the actual value and the applied value can be represented by the following equation:

if the FI-IQMM does not change too much, the error may not be large and performance may be acceptable.

If the error is large, the error can be compensated as follows:

and

it is observed that

And

error r_eEstimated as:

from

And

can be obtained as follows

According to one embodiment, given the selected filter tap N and delay D, the present IQMC system may estimate all filter coefficients prior to reception of a normal signal using a pilot-based scheme. In one embodiment, the present IQMC system may estimate the filter coefficients using a static calibration scheme employing a single frequency training signal sequence as described above.

The present IQMC system also uses the estimated filter coefficient values to set all filter taps except for filter tap D. The initial values of the filter taps D may be set in several ways. In one embodiment, the present IQMC system may set the filter tap D value to 0. In another embodiment, the present IQMC system may set the filter tap D values to the estimated values obtained using a pilot-based scheme. The estimated filter taps D from the pilot-based scheme may reduce the convergence time for determining optimal filter coefficients. Here, the filter coefficients may be expressed as

As described below, during reception of a normal signal, the present IQMC system performs adaptive processing only for the filter tap D as described below:

in accordance with one embodiment of the method of the present invention,is converged to at filter tap D

The IQMC system may then determine the correction factor as follows

The present IQMC system may then update all other filter taps as follows:

FIG. 5 shows a flow diagram of a training-based calibration scheme according to one embodiment. According to one embodiment, the present IQMC system provides a training-based calibration scheme for calibrating compensator filters. First, the IQMC system transmits a single-frequency signal at a selected frequency (at 501). The present IQMC system determines the frequency response of the corrupted signal at the original frequency and the corresponding image frequency (at 502). The present IQMC system determines an estimate of the frequency response of the compensation filter based on the frequency responses at the original frequency and the mirror frequency (at 503). The present IQMC system repeats the calibration steps performed at 501-503 by sweeping a single frequency in multiple steps (multiple subcarrier spacings) to obtain a snapshot of the frequency response of the compensation filter (at 504). The present IQMC system converts the frequency response of the compensation filter to determine the optimal time domain filter taps for the compensation filter (at 505). For example, the present IQMC system performs a pseudo-inverse of the time-frequency transform matrix to perform a frequency-domain to time-domain transform. The present IQMC system repeats the static calibration steps performed in 501-505 by applying different delays with the same set of single-frequency signals, and checks the frequency response of the corrupted signals (at 506). The present IQMC system then determines the optimal delay that provides the smallest Least Squared Error (LSE) for the corresponding time domain filter tap (at 507). The present training-based calibration scheme can effectively compensate for FD-IQMM mismatch and imbalance and can apply the same baseband filter on different channels for a given frequency band. Further, the training-based calibration scheme may compensate for FI-IQMM mismatch and imbalance.

The present IQMC system may also compensate for FI-IQMM that changes with the channel due to changes in the mixer settings. Prior to the reception of a normal signal, the present IQMM system estimates the filter coefficients for a plurality of filter taps N based on the training-based calibration scheme described above. It is understood that the present IQMC system may use other pilot-based static calibration schemes without departing from the scope of the present disclosure. The present IQMC system sets all N filter taps (which do not change with frequency band/channel) using corresponding estimated filter coefficients, sets an initial value of a delay tap D based on a value of 0 or an estimated value obtained using a pilot-based static calibration scheme, and estimates the filter coefficients of the delay tap D (which change with frequency band/channel) during reception of a normal signal according to an iterative scheme based on an adaptive filter. The present IQMC system may also use the converged filter coefficients of the delay tap D to improve the estimated filter coefficients of all N filter taps.

According to some embodiments, the present IQMC system uses a real filter. Fig. 6 illustrates an example of a real filter according to one embodiment. A real filter may apply different filtering schemes on the Q path and the I path. u. u _i(t) and u_q(t) is a pair of impairment signals. The corrupted signal is the partially compensated signal and y (t) is the finally compensated signal. The compensation takes place in two stages. The first stage is used to compensate for the gain mismatch part of FD-IQMM, and FI-IQMM. The second stage is used to compensate the phase mismatch part of the FI-IQMM. The real filter includes a real scaling factor that feeds a delayed version of the in-phase (I) path into the output of the real filter of the quadrature (Q) path.

With appropriate adjustments to the filtering scheme, the performance of the real filter 600 can be matched to that of a similarly configured complex filter. In addition, the real filter 600 has a simpler structure than a complex filter. The simple structure of the real filter can reduce the gate count of the compensator blocks.

Fig. 7 shows a block diagram of an example complex IQMC system, in accordance with one embodiment. The functional blocks of complex IQMC system 700 include: a first complex filter 711 on the main path, a second complex filter 713 on the feedthrough path, an operation block 712 operating on an input signal z (f), an adder 716 adding the signals output from the first complex filter 711 and the second complex filter 713. The mapping of the real and complex compensators is based on a mathematical relationship between the corrupted signal and the compensated signal. In a practical implementation, the two compensators may be completely different. Fig. 1 shows a complex compensator, while fig. 6 shows a real compensator.

Complex IQMC system 700 may be implemented as a signal receiver in a wireless communication system. The complex IQMC system 700 is represented in terms of the frequency domain. For example, the mismatch signal (also referred to herein as the impairment signal) is represented as z (f) as an input to produce a compensated signal y (f). The present complex IQMC system 700 is represented as a complex replica IQMC system 100 shown in fig. 1 comprising a complex filter 711 and a complex filter 713. Operation block 712 is a complex conjugate unit that obtains the complex conjugate of the input signal z (f).

In the time domain, the discrete compensated signal at the output of the real compensator is represented as:

in the frequency domain, the compensated signal is represented as:

equation 27 shows a mathematical model of CVC and is able to calculate IRR of CVC. Using equation 27, RVC can be expressed as an equivalent CVC. After the RVC is written in a form equivalent to the CVC, the IRR of the RVC can be calculated in an analog manner as discussed above. The inputs and outputs of the RVC can be written as analog formsFormula (II) is shown. W₁(f) And W₂(f) Is a function of the parameters used in the RVC. The IRR of RVC can also be calculated using equation 27.

Fig. 8 shows a block diagram of an example IQMC system, according to one embodiment. IQMC system 810 comprises: a signal generator 811, IQ compensation logic 812, and an IQ compensator 813. According to one embodiment, the IQ compensator 813 may implement a delay block and a compensation filter comprising a plurality of time domain filter taps as shown in fig. 1. The IQ compensation logic may be implemented using a hardware digital signal processing chip or implemented in firmware of a receiver in a wireless communication system.

According to one embodiment, the IQ compensation logic 812 provides static calibration of the IQ compensator 813. The signal generator 811 is configured to generate a single frequency signal. According to one embodiment, the IQ compensation logic 812 determines the frequency of the single frequency signal and instructs the signal generator 811 to transmit the single frequency signal at the selected frequency. The signal generator 811 may transmit a series of single-frequency signals to the IQMC by scanning frequencies within a predetermined frequency range. IQ compensation logic 812 receives the impairment signal in response to each single-frequency signal in the series of single-frequency signals, determines an estimate of the frequency response of the compensation filter, and converts the estimated frequency response of the compensation filter into a time domain response to determine optimal time domain filter taps for the compensation filter. The IQ compensation logic 812 also analyzes the overall frequency response of the corrupted signal over a predetermined frequency range. Using the overall frequency response of the corrupted signal, the IQ compensation logic 812 determines the optimal time delay for the delay block and the coefficients of the filter taps of the compensation filter of the IQ compensator 813. For example, the IQ compensation logic 812 generates a time-frequency transform matrix and converts the frequency response of the compensation filter into a time-frequency response using a pseudo-inverse of the time-frequency transform matrix.

According to an embodiment, the IQ compensation logic 812 provides adaptive calibration of the IQ compensator 813 to optimize the time delay of the delay block and the filter taps of the compensation filters of the IQ compensator 813. The IQ compensation logic 812 repeats the static calibration of the IQ compensator 813 with the same set of single frequency signals using multiple delay values. The IQ compensation logic 812 determines an optimal delay among the plurality of delay values that provides a minimum Least Squares Error (LSE).

According to one embodiment, a method for providing IQ mismatch (IQMM) compensation comprises: transmitting a single frequency signal at an original frequency; determining a first response of the corrupted signal at the original frequency and a second response of the corrupted signal at the corresponding image frequency; determining an estimate of a frequency response of the compensation filter at the original frequency based on the first response and the second response; determining a snapshot of a frequency response of the compensation filter by repeating the steps of transmitting a single frequency signal, determining a first response and a second response, and determining an estimate of the frequency response of the compensation filter by sweeping the single frequency signal through a plurality of frequency steps having an interval corresponding to a plurality of subcarrier frequencies of the original frequency; converting the frequency response of the compensation filter into a plurality of time domain filter taps of the compensation filter by performing a pseudo-inverse of a time-frequency transform matrix; the time delay that provides the smallest time domain filter tap is determined based on a plurality of Least Squared Errors (LSEs) for the respective LSE.

The time-frequency transform matrix may be obtained by performing a discrete-time fourier transform (DFT) on a snapshot of the frequency response of the compensation filter.

The method may further comprise: a finite number of filter taps are selected among the plurality of time domain filter taps.

The limited number of filter taps may include one or more positive filter taps.

The method may further comprise: an additional time delay is added in the feedforward path of the compensation filter to include negative filter taps.

The compensation filter may be a complex filter.

The compensation filter may be a real filter comprising a real scaling factor, wherein the real scaling factor feeds a delayed version of an in-phase (I) path into an output of the real filter of a quadrature (Q) path.

The compensation filter may be implemented in a receiver of a wireless communication system.

The compensation filter may comprise a baseband digital filter.

The method may further comprise: estimating filter coefficients for a plurality of time domain filter taps prior to reception of a normal signal based on a static calibration scheme; setting each time domain filter tap of the plurality of time domain filter taps using the respective estimated filter coefficient; setting an initial value of a time delay tap to zero or based on an estimated value obtained using a static calibration scheme; the filter coefficients of the time delay taps are estimated using a normal signal based on an iterative scheme using an adaptive filter.

The IQMM may be a frequency dependent IQMM (FD-IQMM) or a frequency independent IQMM (FI-IQMM).

The compensation filter may be a complex filter.

The compensation filter may comprise a baseband digital filter.

According to one embodiment, an apparatus comprises: a signal generator for generating and transmitting a single frequency signal at an original frequency; a compensator comprising a time delay and a plurality of time domain filter taps; and compensation logic to perform a static calibration of the compensator. The compensation logic is configured to: determining a first response of the corrupted signal at the original frequency and a second response of the corrupted signal at the corresponding image frequency; determining an estimate of the frequency response of the compensation filter at the original frequency based on the first response and the second response; repeating the steps of transmitting a single frequency signal, determining a first response and a second response, and determining an estimate of the frequency response of the compensation filter by sweeping the single frequency signal through a plurality of frequency steps having an interval corresponding to a plurality of subcarrier frequencies of the original frequency to determine a snapshot of the frequency response of the compensation filter; converting the frequency response of the compensation filter into a plurality of time domain filter taps of the compensation filter by performing a pseudo-inverse of a time-frequency transform matrix; the time delay that provides the smallest time domain filter tap is determined based on a plurality of Least Squared Errors (LSEs) for the respective LSE.

The compensation logic may be further configured to: a finite number of filter taps are selected among the plurality of time domain filter taps.

The finite number of filter taps may include one or more positive filter taps.

The compensation logic may be further configured to: an additional time delay is added in the feed-forward path of the compensation filter to include negative filter taps.

The compensation filter may be a complex filter.

The compensation logic may be further configured to: estimating filter coefficients for a plurality of time domain filter taps prior to reception of a normal signal based on a static calibration scheme; setting each time domain filter tap of the plurality of time domain filter taps using the respective estimated filter coefficient; setting an initial value of a time delay tap to zero or based on an estimated value obtained using a static calibration scheme; the filter coefficients of the time delay taps are estimated using the normal signal based on an iterative scheme using an adaptive filter.

The above-described example embodiments have been described above to illustrate various embodiments to implement systems and methods that provide IQ mismatch calibration and compensation. Various modifications and alterations from the disclosed example embodiments will become apparent to those skilled in the art. The subject matter that is intended to be within the scope of this disclosure is set forth in the claims.

Claims

1. A method for providing in-phase quadrature mismatch compensation, the method comprising:

transmitting a single frequency signal at an original frequency;

determining a first response of the corrupted signal at the original frequency and a second response of the corrupted signal at the corresponding image frequency;

determining an estimate of the frequency response of the compensation filter at the original frequency based on the first response and the second response;

determining a snapshot of a frequency response of the compensation filter by repeating the step of transmitting a single frequency signal, the step of determining a first response and a second response, and the step of determining an estimate of the frequency response of the compensation filter by sweeping the single frequency signal through a plurality of frequency steps, wherein the plurality of frequency steps have an interval corresponding to a plurality of subcarrier frequencies of the original frequency;

converting the frequency response of the compensation filter into a plurality of time domain filter taps of the compensation filter by performing a pseudo-inverse of a time-frequency transform matrix;

The time delay providing the smallest least squared error is determined based on a plurality of least squared errors for respective time domain filter taps.

2. The method of claim 1, wherein a time-frequency transform matrix is obtained by performing a discrete-time fourier transform on a snapshot of the frequency response of the compensation filter.

3. The method of claim 1, further comprising: a finite number of filter taps are selected among the plurality of time domain filter taps.

4. The method of claim 3, wherein the finite number of filter taps comprises one or more positive filter taps.

5. The method of claim 4, further comprising: an additional time delay is added in the feed-forward path of the compensation filter to include negative filter taps.

6. The method of claim 1, wherein the compensation filter is a complex filter.

7. The method of claim 1, wherein the compensation filter is a real filter comprising a real scaling factor, wherein the real scaling factor feeds a delayed version of an in-phase path into an output of the real filter of a quadrature path.

8. The method of claim 1, wherein the compensation filter is implemented in a receiver of a wireless communication system.

9. The method of claim 1, wherein the compensation filter comprises a baseband digital filter.

10. The method of claim 1, further comprising:

estimating filter coefficients of the plurality of time domain filter taps prior to reception of a normal signal based on a static calibration scheme;

setting each time domain filter tap of the plurality of time domain filter taps using the respective estimated filter coefficient;

setting an initial value of a time delay tap to zero or based on an estimated value obtained using a static calibration scheme;

the filter coefficients of the time delay taps are estimated using a normal signal based on an iterative scheme using an adaptive filter.

11. An apparatus for providing in-phase quadrature mismatch compensation, the apparatus comprising:

a signal generator for generating and transmitting a single frequency signal at an original frequency;

a compensator comprising a time delay and a plurality of time domain filter taps;

compensation logic to perform a static calibration of the compensator,

wherein the compensation logic is configured to:

Determining an estimate of a frequency response of the compensation filter at the original frequency based on the first response and the second response;

determining a snapshot of a frequency response of the compensation filter by repeating the steps of transmitting a single frequency signal, determining a first response and a second response, and determining an estimate of the frequency response of the compensation filter by sweeping the single frequency signal through a plurality of frequency steps having an interval corresponding to a plurality of subcarrier frequencies of the original frequency;

converting a frequency response of the compensation filter into a plurality of time domain filter taps of the compensation filter by performing a pseudo-inverse of a time-frequency transform matrix;

12. The device of claim 11, wherein the compensation logic is further configured to: a finite number of filter taps are selected among the plurality of time domain filter taps.

13. The apparatus of claim 12, wherein the finite number of filter taps comprises one or more positive filter taps.

14. The device of claim 13, wherein the compensation logic is further configured to: an additional time delay is added in the feed-forward path of the compensation filter to include negative filter taps.

15. The apparatus of claim 11, wherein the compensation filter is a complex filter.

16. The apparatus of claim 11, wherein the compensation filter is a real filter comprising a real scaling factor, wherein the real scaling factor feeds a delayed version of an in-phase path into an output of the real filter of a quadrature path.

17. The device of claim 11, wherein the compensation logic is further configured to: